Recommendation for fertilizer application for soils via qualitative reasoning

In Brazil, liming and fertilization are essential practices in agriculture due to the soils being acidic and poor in nutrients. Distinct decision tables that serve as support for recommendation of fertilizer application are used, but one of their features is that they are based on a strictly quantitative analysis of input variables. This sometimes causes di®culties in their use when calculating the output (recommended fertilizer application). This work presents a model for the use and interpretation of decision tables for fertilizer application. It is based on a qualitative characterization for the rules and input variables used. The results have shown that this approach gives feasible results which more accurately re ̄ect human thinking about the decision table. # 2001 Elsevier Science Ltd. All rights reserved.